mgt 560 queuing system simulation stochastic modeling © victor e. sower, ph.d., c.q.e. 2007
TRANSCRIPT
MGT 560Queuing System Simulation
Stochastic Modeling
© Victor E. Sower, Ph.D., C.Q.E. 2007
Steps in Simulation Process
1. Define problem2. Define important variables in problem3. Collect data4. Construct mathematical model5. Validate model6. Define experiments to run7. Run experiments8. Consider results (possible model modification)9. Decide on course of action
Victor E. Sower, Ph.D., C.Q.E. 2007
Advantages of Simulation
• Straightforward and flexible• Can analyze complex real-world situations• Can use any distributions—not just standard ones• Time compression• Can address “what-if” questions• Off-line• Can study interactions of individual variables and
components
Victor E. Sower, Ph.D., C.Q.E. 2007
Limitations of Simulation
• Expensive and time consuming• Does not generate optimal solutions• The results from the model are limited by the
quality of the design of the model• Each simulation model is unique to a particular
problem
Victor E. Sower, Ph.D., C.Q.E. 2007
Types of Queuing Systems
Single channel; Single phase
Victor E. Sower, Ph.D., C.Q.E. 2007
Channel – the number of parallel servers
Phase – the number of servers in sequence
Types of Queuing Systems
Multiple channel; Single phase
Victor E. Sower, Ph.D., C.Q.E. 2007
Types of Queuing Systems
Single channel/Multiple phase
Victor E. Sower, Ph.D., C.Q.E. 2007
Types of Queuing Systems
Multiple channel/Multiple phase
Victor E. Sower, Ph.D., C.Q.E. 2007
Data Collection
• Source of customers– Infinite– Finite
Victor E. Sower, Ph.D., C.Q.E. 2007
Data Collection
• Arrival Rate/Interarrival Time– Arrival Rate (Poisson)– Interarrival Time (Exponential)
Victor E. Sower, Ph.D., C.Q.E. 2007
Data Collection
• Service Rate/Service Time– Service Rate (Poisson)– Service Time (Exponential)
Victor E. Sower, Ph.D., C.Q.E. 2007
Data Collection• Queue Discipline– FCFS– LIFO– Random– Others
Victor E. Sower, Ph.D., C.Q.E. 2007
Data Collection
• Queue Length– Infinite– Finite• Balking
Victor E. Sower, Ph.D., C.Q.E. 2007
System Operating CharacteristicsResults from Model
• L Avg. no. of customers in system• Lq Avg. no. of customers in the queue• W Avg. time customer spends in system• Wq Avg. time customer spends in queue• p Utilization rate
Victor E. Sower, Ph.D., C.Q.E. 2007
System Considerations
• Waiting line costs• Service quality• Psychology of waiting• Balking
Victor E. Sower, Ph.D., C.Q.E. 2007